Re: [PATCH] PR debug/45682 - wrong struct DIE nesting with -fdebug-types-section

Cary Coutant writes: > Take a look at this and see if it makes sense; Yes, it does make sense to me. Thanks. -- Dodji

Re: [Patch, Fortran] PR51578 Fix symbol import with renaming and only

Hi Tobias, Your patch passed my tests and in top of that fixes the ICE for pr51522.f90. However I don't understand the errors: [macbook] f90/bug% gfc -w pr51522.f90 pr51522.f90:268.10: integer(c_int) function vect_Utils_vuTriIneqHolds_c(u, v, tol, exception)& 1 Error: Parameter '

Re: [Patch, fortran] PR51791 - [OOP] Failure to resolve typebound function call with base object in parentheses

Dear Paul, Paul Richard Thomas: A question for the standard aficianados: Are there other base object expressions that are legal? I don't think so. (Ignoring RESHAPE, SPREAD etc., cf. PR 47505.) Bootstrapped and regtested on FC9/x86_64 - OK for trunk? OK. Thanks for the patch. ***

Re: [Patch, Fortran] PR51652 - alloc with type-spec: check that char len matches declaration

Please ignore the following part of my patch - its unrelated. (It's for/from http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46328#c11) Tobias Index: gcc/fortran/resolve.c === --- gcc/fortran/resolve.c (Revision 182995) +++ gcc/

[Patch, Fortran] PR51652 - alloc with type-spec: check that char len matches declaration

Rather simple patch. Build and regtested on x86-64-linux. OK for the trunk? Tobias 2012-01-08 Tobias Burnus PR fortran/51652 * resolve.c (resolve_allocate_expr): For non-deferred char lengths, check whether type-spec matches declaration. 2012-01-08 Tobias Burnus PR fortran/51652 *

[Patch, Fortran] PR51578 Fix symbol import with renaming and only

Hi all, gfortran mishandles the module import, if one uses multiple USE statements and rename/only. This part got "broken" when the following has been implemented: "More than one USE statement for a given module may appear in a specification part. If one of the USE statements is without an O

[RFC, patch] libitm: Filter out undo writes that overlap with the libitm stack.

The protection of the libitm stack from undo/redo writes got lost for all the write-through TM methods and logging of thread-local data. I revived the code that we had before for the gtm_cacheline-based writes, but it might have been broken before. The problem is that we don't know where the bott

PR tree-optimization/51680 (indirect inlining in templates)

Hi, this patch fixes second part of problem in PR51680 and moves it out of regression land. The PR is still valid about fact that we should inline process_fun_at even when it is called many times because further inlining of add1 simplifies it enough. Unfortunately I don't think this can be easil

[Patch, fortran] PR51791 - [OOP] Failure to resolve typebound function call with base object in parentheses

Dear All, Having stated in the PR that I did not have time to fix it, after a few hours in the workshop doing woodwork I alighted on the obvious and simple solution :-) A question for the standard aficianados: Are there other base object expressions that are legal? Clearly this fix is extendable.

PR tree-optimize/51694

Hi, in the testcase bellow there is the usual problem with mismatched function arguments. I've added bounds check to avoid the ICE. Bootstrapped/regtested x86_64-linux. Will commit it shortly. Honza PR tree-optimize/51694 * ipa-cp.c (ipa_get_indirect_edge_target): Add bounds check

[PATCH] PR51471. Don't speculate frame-related insns during delayed branch scheduling.

Jeff, this patch fixes PR51471. It prevents frame-related insns from being speculated by delayed branch scheduling, which may cause inconsistencies in cfi state. For one of the examples from PR51471: 20040811-1.c, the scenario is as follows. Before delayed branch scheduling, 20040811-1.c.220r.va

[PATCH] [committed obvious] Fix small leak in cfgloop.c

Hi, I noticed that cfgloop.c:cancel_loop() is missing a free of the bbs array obtained from get_loop_body(). Committed attached patch as obvious. Thanks, Chung-Lin 2012-01-08 Chung-Lin Tang * cfgloop.c (cancel_loop): Add free() of bbs array. Index: cfgloop.c ==

Re: [patch] docs for __STRICT_ANSI__ should mention C++

On Sun, 8 Jan 2012, Jonathan Wakely wrote: > 2012-01-08 Jonathan Wakely > > * doc/cpp.texi (Predefined Macros): __STRICT_ANSI__ is defined > for -std=c++98 and -std=c++11 too. > > OK for trunk? OK. -- Joseph S. Myers jos...@codesourcery.com

[patch] docs for __STRICT_ANSI__ should mention C++

2012-01-08 Jonathan Wakely * doc/cpp.texi (Predefined Macros): __STRICT_ANSI__ is defined for -std=c++98 and -std=c++11 too. OK for trunk? Index: doc/cpp.texi === --- doc/cpp.texi(revision 182988) +++ doc/c

[PATCH] PR51271. Fix handling of frame-related insn in delay slot of annulled branch

Richard, this patch fixes PR51271. It fixes the handling in scan_trace of an insn in the delay slot of an annulled branch that annuls the insn when the branch is not taken. The way the insn is handled, is that: - the insn is applied to the state - the updated state is registered at the jump targ

add pretty printer for std::forward_list

2012-01-08 Jonathan Wakely * python/libstdcxx/v6/printers.py (StdForwardListPrinter): Add. * testsuite/libstdc++-prettyprinters/cxx11.cc: New. Tested 'make check RUNTESTFLAGS=prettyprinters.exp' and committed to trunk The test also checks some of the other C++11 containers' pr

[patch middle-end]: Fix PR/48814 - [4.4/4.5/4.6/4.7 Regression] Incorrect scalar increment result

Hi, this patch makes sure that for increment of postfix-increment/decrement we use also orignal lvalue instead of tmp lhs value for increment. This fixes reported issue about sequence point in PR/48814 ChangeLog 2012-01-08 Kai Tietz PR middle-end/48814 * gimplify.c (gimp

Re: [PATCH] MIPS16 TLS support for GCC

Chung-Lin Tang writes: > Hi Richard, > here are the patches for MIPS16 TLS. Thanks for all the work. It's great to see this hole in the MIPS16 support finally being plugged. There seem to be three changes in the patch: (1) Enable MIPS16 TLS (which must of course work independently of wheth